Elsevier

Small Ruminant Research

Volume 203, October 2021, 106489
Small Ruminant Research

Inferring causal relationships among growth curve traits of Lori-Bakhtiari sheep using structural equation models

https://doi.org/10.1016/j.smallrumres.2021.106489Get rights and content

Highlights

  • Five non-linear models were compared and the Brody model was the best non-linear for describing the growth curve.

  • Standard, fully recursive and temporal recursive models were compared for genetic evaluation of growth curve traits.

  • The structural coefficients of birth weight on b, k and A were estimated −0.187, 0.004 and 12.575, respectively.

  • Causal effects estimates showed that any improvement in initial body weight has a positive effect on their latter weights.

Abstract

This study aimed to infer the causal effects of birth weight on growth curve traits as well as, assessment of causal relationships among the growth curve traits in Lori-Bakhtiari sheep using structural equation modeling. The body weight-age records were collected from birth to 390 days of age from 1995 to 2012 in Shooli Breeding Station of Lori-Bakhtiari sheep, Shahr-e Kord, Iran. Initially, five non-linear models including Brody, Logistic, Negative exponential, Gompertz and von Bertalanffy were fitted on 22,546 body weight-age records for determining the best model describing the growth curve. Among the tested non-linear growth models, Brody model showed the best goodness of fit. Then, data on growth curve traits of 3,168 lambs from 217 rams and 1,211 ewes were considered for genetic evaluation of growth curve traits including asymptotic weight (A), integration constant related to initial animal weight (b), and maturation rate (k).The causal effect of birth weight (BW) on growth curve traits under two models including standard multivariate model (SMM) and fully recursive model (FRM) were fitted via a Bayesian approach. The posterior means of heritability for the growth curve traits were similar under both SMM and FRM, while the posterior means of genetic correlations among the traits were statistically different under both SMM and FRM. Structural coefficients related to causal effects of BW on b, k and A, were statistically significant with values −0.19, 0.01and 12.58, respectively. Therefore, considering the causal effects of BW on growth curve traits is essential for accurate genetic analysis of growth curve traits of Lori-Bakhtiari sheep.

Introduction

Lori-Bakhtiari sheep is one of the most important Iranian native sheep breeds in southwestern of Iran, with a population of more than 1.7 million heads (Vatankhah et al., 2019). This breed is more pronounced for meat production and lambs with fast growth especially from birth to six-month of age.

Growth is one of the most important characteristics of livestock animals (Bathaei and Leroy, 1998; Keshkin et al., 2010; Ghavi Hossein-Zadeh, 2015c, 2017) and can be described as an increase in live bodyweight over time. Moreover, changes in live bodyweight across time are explained by the growth curves (Keshkin et al., 2010). Analysis of the growth performance of animals throughout their lifetime is a valuable tool for administration of the appropriate feeding method and selection for attaining the best slaughter age (Ghavi Hossein-Zadeh, 2015c). Early body weights especially birth weight had a significant effect on latter body weights. Growth curve parameters may be used as criteria for changing the association between live body weight and age via selection and breeding (Kachman and Gianola, 1984). Furthermore, the growth curve parameters are heritable in different species and it may be possible to change the shape of growth curve through selection of animals for those parameters (Kachman et al., 1988). Slow growth rate resulting in low market weight has been identified to be one of the factors limiting profitability in any production system of animals (Noor et al., 2001; Abegaz et al., 2010).

The application of mathematical growth models provides an efficient method for summarizing the information included in a dataset into a few parameters to facilitate both the understanding of the growth trajectory and its interpretation (Malhado et al., 2009). In biological systems, phenotypic traits may have causal relationships but standard multivariate models (SMMs), which are used commonly for genetic evaluation purposes, cannot handle such causal relationships and only estimate association among the traits, while association cannot show causal relationships among the traits. Structural equation models (SEMs) consider simultaneous and/or recursive causal relationships among the traits (Gianola and Sorensen, 2004).

Genetic parameters of growth curve traits in several sheep breeds have been reported using SMMs (Ghavi Hossein-Zadeh, 2015b, 2017; Nimase et al., 2017). These studies did not take the potential causal relationships among the growth curve traits into account. A few numbers of studies have investigated the causal relationships between growth traits in small ruminants (Mokhtari et al., 2018; AmouPosht-e Masari et al., 2019), but these studies did not take into account the growth curve parameters. Growth curve parameters are related to rate of maturing and mature weight; furthermore, these traits have been suggested to have an association with lifetime productivity parameters (Pala et al., 2005). Finding the causal relationships between birth weight and growth curve parameters provides a sound basis for developing a breeding strategy to modify the trajectory of growth (Abegaz et al., 2010). Therefore, the present study aimed to infer the potential causal effects of birth weight on growth curve traits in Lori-Bakhtiari sheep using SEM, as well as to assess the potential causal relationships among the growth curve traits. Besides, estimates of genetic parameters, goodness of fit and predictive ability of SMM and SEM were compared.

Section snippets

Flock management

The flock was managed under a semi-migratory or village system. From December to May the flock was usually kept at the breeding station and was fed lucerne, barley, and wheat stubble indoors. The rest of the year the flock grazed on the range and cereal pasture. The mating period extended from late August to late October (20–25 ewes were assigned randomly to one fertile ram) and consequently, lambing started in late January and lasted to late March. Lambs were suckling from their dams until 15

Results and discussion

The goodness of fit statistics including R2adj, AIC, RMSE and DW for comparing the non-linear models describing the growth trajectory of Lori-Bakhtiari lambs are presented in Table 3.The Brody model supplied the highest values of R2adj (0.89) and DW (1.62) and the lowest values of AIC (149895.68). The estimates of growth curve parameters using the non-linear models are shown in Table 4.

The BIC values across the traits under six different animal models are shown in Table 5. The best animal model

Posterior means of direct heritability estimates

Posterior means of direct heritability of BW under SMM (0.32) and FRM (0.31) were in line with those reported by others (Safari et al., 2005; Tariq et al., 2010). Under both SMM and FRM, the estimates of direct heritability for b, k and A were 0.14, 0.16 and 0.10, respectively. Hence, the existence of potential causal relationships among the studied traits did not have significant effect on the direct heritability estimates of studied traits. Ghavi Hossein-Zadeh (2015a) reported values of 0.15,

Conclusions

The obtained results showed that the differences between direct heritability estimates from SMM and FRM were negligible, while correlation estimates between the studied traits using SMM and FRM were statistically significant. The causal effect of BW on k and A, showed that any improvement in the initial weight of Lori-Bakhtiari lambs has a positive effect on their latter weights. Therefore, BW is an important trait for improving the growth performance of Lori-Bakhtiari breed. However,

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Declaration of Competing Interest

The authors report no declarations of interest.

Acknowledgment

The authors would like to thank all the staff of Lori-Bakhtiari Shooli Station for data collection and the National Animal Breeding Center (ABC) of Iran for providing the data.

References (37)

  • y. Akbas et al.

    Comparison of several models to fitting growth curves of Kivircik and Daglic male lambs

    Turk. J. Vet. Anim. Sci.

    (1999)
  • H. AmouPosht-e Masari et al.

    The comparison of alternative models for genetic evaluation of growth traits in Lori-Bakhtiari sheep: implications on predictive ability and ranking of animals

    Small Rumin. Res.

    (2019)
  • S. Brody

    Bioenergetics and Growth With Special Reference to the Efficiency Complex in Domestic Animals

    (1945)
  • J.E. Brown et al.

    A comparison of nonlinear models for describing weight-age relationships in cattle

    J.Anim. Sci.

    (1976)
  • J. Durbin et al.

    Testing for serial correlation in least square regression

    Biometrika.

    (1950)
  • A.B. Gbanboche et al.

    Comparison of non-linear growth models to describe the growth curve in West African Dwarf sheep

    Anim.

    (2008)
  • N. GhaviHossien-Zadeh

    Bayesian estimates of genetic relationship between growth curve parameters in Shall sheep via Gibbs sampling. Iran

    J. Appl. Anim. Welf. Sci.

    (2015)
  • N. GhaviHossien-Zadeh

    Estimation of genetic relationships between growth curve parameters in Guilan sheep

    J. Anim. Sci. Technol.57

    (2015)
  • Cited by (0)

    View full text